In the Trans-Himalaya, the Kohistan–Ladakh-Karakoram Batholiths have evolved as a single entity during the late Cretaceous and early Palaeogene (Pundir et al., 2020, and references therein). The Ladakh Batholith represents an integral part of the Trans-Himalayan calc-alkaline magmatic belt extending from east of Nanga Parbat to Lhasa (Figure 1A) and is bounded by Shyok Suture Zone (SSZ) in the north and ITSZ in the south (Figure 1B). The Asian plate in the north and Indian plate in the south are juxtaposed and accreted along the ITSZ with the closure of Neo-Tethyan Ocean. Ahmad et al. (1998) stated that the enclaves in the granites probably represent the initial pulses of magmatism in response to intra-oceanic northward subduction of Neo-Tethyan Ocean beneath an immature arc, and subsequently, a huge amount of felsic magmatism was generated as the arc matured forming the Ladakh Batholith.

The Ladakh Batholith is differentially unroofed in the western (p = 3.35–4.27 kbar), central (p = 2.99 kbar), and eastern (p = 2.17 kbar) parts, as inferred from Al-in-hornblende geobarometer (Kumar, 2010b). The FeO^t/MgO ratio of biotites from the granites of western (1.44–2.19), central (1.31–1.90), and eastern (1.71–2.63) parts suggests the contribution of dominant calc-alkaline, metaluminous (I-type), and subordinate amount of peraluminous (S-type) felsic magmas in the evolution of Ladakh Batholith. Ahmad et al. (1998) observed that the enclaves and mafic dykes in granites of the Ladakh Batholith are basaltic-andesite and andesite belonging to subalkaline, calc-alkaline series in terms of major and trace elements. The negative relationship of MgO, Fe₂O₃ ^t, MnO, TiO₂, CaO, Al₂O₃ against SiO₂, and positive trends of TiO₂, CaO, CaO/Al₂O₃ against MgO may indicate fractionation of gabbroic assemblage, however, the absence of negative chondrite normalized Eu and primitive mantle normalized Sr anomalies shows that plagioclase was not an important fractionating phase. Based on geochemical features of enclaves, Ahmad et al. (1998) proposed the cognate or autolith origin for some selected enclaves sampled from western Ladakh, as inferred by Honegger et al. (1982), and others as entrained xenoliths of pre-existing mafic rock into the granites. However, Kumar and Singh (2008) suggest that the enclaves and host granites are largely as calc-alkaline, mildly to strongly metaluminous (molar Al₂O₃/CaO + Na₂O + K₂O = 0.9–1.05), which are compositionally different from diorite, gabbro, and mafic dykes. Near linear to curvilinear variation trends observed for TiO₂, CaO, Fe₂O₃ ^t, MgO, MnO, and TiO₂ against SiO₂ of enclaves and host granites can be attributed to synchronous mafic-felsic magma mixing and fractionation. The large data scatter for Al₂O₃, Na₂O, K₂O, P₂O₅, Rb, Ba, Nb, Yb, and Zr against SiO₂ is, however, a combined outcome of elemental diffusion at varying degrees and sorting of minerals hosting these elements during mixing-fractionation events.

The extent and duration of calc-alkaline felsic magmatism constituting the Ladakh Batholith have been constrained based on a variety of chronological database, without taking into account the mafic to hybrid magma contributions. Magmatism within the Ladakh Batholith was most intense during Late Cretaceous to Early Paleogene (67 and 45 Ma) (Singh et al., 2007; Upadhyay et al., 2008; Jain and Singh, 2009; Ravikant et al., 2009; White et al., 2011a; White et al., 2011b; Kumar et al., 2012; Shellnutt et al., 2013; Jain 2014) with early discrete pulses at ca. 102 and 97 Ma identified in the western part (Honegger et al., 1982). Some sporadic geochronological data (Singh et al., 2007) on coarse grained diorite (58.4 ± 1.0 Ma) and granodiorite (60.1 ± 0.5 Ma) from Karu and Chang La regions of the Ladakh Batholith suggest a coeval nature between the diorite and granodiorite. A subvolcanic mafic dyke (45.7 ± 0.8 Ma) and granites (45.27 ± 0.56 Ma) at Daah-Hanu (Weinberg and Dunlop, 2000) provide the same ages of their formation. The likely reasons for synchronous diorite and granodiorite magmatism are not yet described. However, in and around the Leh and Sabu regions, stock-like dioritic bodies are also found as pre-existing older lithological units as compared to the granites (Kumar et al., 2016; Weinberg, 1997).

The modal mineralogy of the studied granites represents quartz diorite, granodiorite, and monzogranite whereas enclaves are micro-monzodiorite, micro-quartz diorite, micro-tonalite, and micro-monzogranite. The granites and enclaves have a common mineral assemblage (hbl-bt-pl-Kf-qtz-ap-zrn-mag ± ilm; mineral symbols are after Kretz, 1983) but they differ in modal proportion, the enclaves being enriched in mafic minerals.

Microscopically, the enclaves exhibit typical magmatic textures such as fine grained, equigranular, hypidiomorphic to porphyritic (Figures 8, 9). The contacts between the enclaves and host granites are sharp and wavy without any reaction signature (Figures 8A,B). In some instances, the contact outline between them is marked by the grain boundaries of either plagioclase or poikilitic amphibole crystallized in the host granite (Figures 8A,B). The grain boundaries of these minerals are devoid of any thermal effect and probably served as surfaces against which the enclave magma globules might have undercooled uniformly, and hence produced fine grained equigranular texture. In some enclaves, biotite dominates over the amphiboles and vice versa (Figures 8A,B). The presence of magnetite both in the enclave and granite indicates oxidizing condition, which is a characteristic of calc-alkaline subduction zone magmatism. The small flaky and platy biotites are commonly aligned along the contact outline (Figure 8A), and hence could not penetrate the semi-solid enclaves during the mingling event.

In a few enclaves, subhedral to euhedral amphiboles along with a minor amount of biotite are clustered and form the mafic aggregates (Figure 8C) that are in close association with tabular plagioclase, interstitial quartz, and K-feldspar. Subhedral to anhedral elongate amphiboles along with magnetite, quartz, and plagioclase show magmatic flowage texture (Figure 8D). Small acicular (needle-shaped) apatite crystals can also be observed poikilitically enclosed within the interstitial felsic minerals, which result from the undercooling of enclave magma globules. Porphyritic enclaves frequently contain plagioclase phenocrysts with normal, patchy, and oscillatory zonings (Figures 9A,B). They exhibit resorption surfaces as a corroded (dissolution) feature over which albitic rims are grown (Figure 9B). In some cases, the albitic rim appears to be replaced by sericite (Figure 9A). Synplutonic mafic dykes and sheets (partly pillow-like) show a typical interlocking subpoikilitic texture where plagioclase laths are partially enclosed by anhedral to subhedral amphiboles and less frequently by biotite, commonly associated with granular magnetite (Figures 9C,D).

Pre-existing stock-like dioritic bodies (Weinberg, 1997) are intruded by the granite-leucogranite-pegmatite system, as evident from the presence of dioritic xenoliths hosted in the intruding lithological units (Figures 2A,B). It is likely that the Neo-Tethyan oceanic subduction-related magmatism may have been initiated with the emplacement of diorite followed by profuse granite magmatism in the Ladakh Himalaya. The rarely occurring layered structure (Figure 2C) in the granite, exposed close to Sabu, is likely due to mafic-felsic mineral sorting under the influence of double diffusion convection and gravity settling operated in a magma chamber (e.g., McBirney, 1985), and not due to the enclave-forming magmatic pulse. The hornblende biotite granite is a major rock constituent of the Ladakh Batholith, which commonly hosts the microgranular enclaves. However, a biotite granite exposure (Figure 2D) contains a fine grained mesocratic enclave with diffused contacts (Figure 2D), which possibly represents cognate enclave of early crystallized fine grained cumulus phases of host magma.

The origin of igneous enclaves hosted in the granites is still debated (Clemens et al., 2018, and references therein). The presence of mafic clots, inherited and metamorphic zircons, and Ca-rich patchy zoned (resorbed) plagioclase in enclaves favors the restite origin of enclaves (e.g., Chappell et al., 1987; Chen et al., 1989; Huang et al., 2018). The studied enclaves from Ladakh Batholith, however, lack partial melting textures, deformational, or recrystallization features. In contrast, the enclaves show typical fine grained to porphyritic textures, acicular apatite, some elongated amphiboles (Figures 8C,D, 9A,B), sharp, and lobate enclave-granite contact outlines (Figures 3–5), and patchy zoned core with a corrosive surface of plagioclase phenocrysts (Figures 9A,B), which together oppose the restite origin of enclaves. The observed anorthite rim (bright gray zone) grown over the corroded margin of plagioclase xenocrysts (Figures 9A,B) likely crystallized in a new mafic-felsic magma mixed (high-T) environment (e.g., Hibbard, 1981; Andersson, 1988; Andersson and Eklůnd, 1994; Kumar et al., 2017). The enclave-bearing granite may have, however, experienced later tectono-thermal deformation (e.g., Siebel et al., 2012).

The cognate model states that the enclave is a cumulate rock of early crystallized cumulus phases or rapidly cooled (chilled) border rock series of the host granite magma or more mafic cogenetic fractions with trapped interstitial melts (e.g., Phillips et al., 1981; Dodge and Kistler, 1990; Dahlquist, 2002; Esna-Ashari et al., 2011; Xiao et al., 2020). The cumulate hypothesis appears valid where the evidence of excessive nucleation of ferromagnesian minerals by rapid cooling at the margins of a shallow magma chamber and later its fragmentation and dispersal by dynamic arc plutonism is preserved (e.g., Esna-Ashari et al., 2011). Similarly, reheating and remelting of the mafic chilled margin in the ascent conduits can be dislocated by dyke injection and the host rocks can be fragmented to sizes similar to the enclaves (e.g., Fernandez and Castro, 2018). However, in the present case, numerous features such as sharp and crenulate contact outline (Figures 3–5), xenocrystic-bearing (hybrid) enclaves (Figures 9A,B), and large grain size contrast between the enclave and host granite (Figures 8A,B) strongly oppose cognate origin of enclaves. Coarse grained plutonic xenoliths have been reported in the volcanic system (Cooper et al., 2019) but in the felsic plutonic system, they represent early formed crystal mush pooled at the base of the magma chamber (e.g., Wiebe et al., 1997).

The mafic-felsic magma interacting system truly represents a composite mafic-silicic intrusive system (Wiebe, 1980; Wiebe, 1994) typical to those observed in the present study.

Most magmas experience three rheological states during crystallization, 1) Newtonian state at a low fraction of crystals (F < 0.35), 2) Bingham behavior or stress thinning state with yield stress (F = 0.35–0.65), and 3) solid-like behavior (F > 0.65) (Fernandez and Gasquet, 1994, and references therein). Coeval mafic and felsic magmas with contrasting temperatures and compositions achieve thermal equilibration at which crystallization is more advanced in the mafic magma rather than the felsic magma. At this point, an inversion temperature exists that governs relative viscosities of magmas such as mafic magma becomes less viscous than felsic magma at a temperature higher than the inversion temperature, and vice versa. Thus, the generated inversion of viscosity between the coeval mafic and felsic magmas can explain most of the features described here (e.g., Frost and Mahood, 1987; Fernandez and Gasquet, 1994). However, an open felsic magma chamber intermittently recharged by mafic to hybrid magma pulses may behave in a different and more complex manner. Magma convection ceases once the magma chamber develops a crystallinity of ca 40–50% (Chen et al., 2018).

Enclaves of variable shapes and sizes are formed within the host granite magma with a low crystallinity at an initial stage of its evolution. The enclaves may lack a chilled margin because small size enclaves (ca. 15 cm) are not capable to generate a thermal gradient within it, and hence undercooled at a uniform rate. The large sized enclave globules may, however, generate a thermal gradient forming the chilled margin and coarser core, because they behave as micro-plutons. It is therefore suggested that each enclave partially crystallizes as a discrete body with a unique cooling history (e.g., Rooyakkers et al., 2018). Numerical modeling of deformation of a crystal-rich enclave in a pure shear regime suggests that enclave particles pack early, quickly erasing differences in initial content and building a force chain parallel to the compression axis that transmits stresses to the host granite (Burgisser et al., 2020). It is observed that the enclaves are not deformed under solid state and display similar magnetic fabrics that imply mafic magma globules coexisted and flowed with felsic host magma (Zhu et al., 2018). The mineral constituents of the studied enclaves of the Ladakh Batholith are solely unaffected by deformation, recrystallization, and solid-liquid reaction rims (Figures 8, 9). Vernon (1983) advocated if the magmatic flowage passes through the enclaves, the homogeneity may approach partial to complete disappearance of enclaves, as observed at some places in the present case (Figures 4A,B). This is a process leading to homogenization under a flow-favoured magma mixing environment (e.g., Reid and Hamilton, 1987; Reid et al., 1983; Castro, 1990).

How are the enclaves with diverse geometry and nature brought together at one place (Figures 3D, 5C) in a felsic magma chamber? The most plausible reason is either whole body convection current in the magma chamber or due to gravity settling at an early stage of granite evolution (e.g., Wiebe et al., 2007). However, over the time scale of magma mixing and mingling, convective patterns are difficult to recognize (Montagna et al., 2015). The absence of mafic schlierens among the randomly distributed enclave population indicates the transportation of enclaves in a much weaker medium (low viscosity) of host granite magma. Igneous lamination and schlieren in granites (Figures 3B, 4A,B) are, however, an expression of concurrent magma injection, aggregation with or without mingling/hybridization, hydrodynamic processes coupled with fractional crystallization, and deformation during the growth of plutons (Barbey, 2009). Thus, within a given volume of Ladakh granites and based on the type and behavior of enclaves, one can recognize the comingling and hybrid zones (e.g., Mariano and Sial, 1988; Perugini et al., 2003). Although enclaves do not provide precise information on the depth (pressure) where the felsic-mafic magmas were mixed, the enclaves appear to have ascended from greater depths than their surrounding host granites (Didier et al., 1989).

Water content, crystal size, degree of crystallinity, mass fraction, composition, and temperature determine the rheology of interacting mafic and felsic magmas (Frost and Mahood, 1987). Low viscosity, minimal rheological differences, and thermal equilibrium between the coeval mafic and felsic magmas will produce convective overturn (Huppert et al., 1984; Sparks and Marshall, 1986) forming a hybrid magma zone hidden below the felsic magma at crustal depth. Under the influence of turbulent convection, the hybrid magma may also be injected into the overlying felsic magma, forming either rounded to elongated hybrid enclave globules or disrupted as angular brecciated to dyke-/sheet-like enclave swarms depending upon the schedule of its injection into evolving felsic magma with changing rheology (e.g., Barbarin, 1989; Fernandez and Barbarin, 1991; Barbarin and Didier, 1992; Fernandez and Gasquet, 1994; Kumar, 2014), as shown by a schematic model (Figure 10). A large fine grained circular composite enclave (Figure 4D) lacks zoning. Zoned enclaves are commonly formed either by internal differentiation mechanism or multiple magma mixing and mingling events from magma source to emplacement level (Castro et al., 1990; Zhang et al., 2014; Zhang et al., 2016) or in two different, deep and shallower, magma chambers (Adam et al., 2019). However, the observed composite enclaves (Figure 4D) appear to have formed in an open felsic magma chamber that was essentially recharged intermittently by mafic to hybrid magmas pooled just below the felsic magma chamber (Figure 11), as similarly suggested for the origin of composite and other types of enclaves found in the granites of western Ladakh Batholith (Kumar, 2010a).

A few sheet-like bodies appear to be pillowed against relatively cooler host granite magma (Figure 6A), similar to as described elsewhere (e.g., Clemens et al., 2016). Mafic synplutonic dykes contain a high volume of crystal content (Figures 9C,D) and hence behaved more like a solid and become brecciated angular enclaves (Figures 7A–C). Depending upon the viscosity of enclave-hosting felsic magma and prevailing local strain, the structures like flowing or static breccias may develop (Fernandez and Gasquet, 1994). Brecciation and disruption (break-up) of mafic to hybrid synplutonic dykes (Figures 6A, 7C) might have occurred within a transition zone between the crystal mush (>65%) and melt condition of host felsic magma (Janoušek et al., 2000; Rooyakkers et al., 2018). However, the hybrid dyke magma might have been sourced from the felsic-mafic hybrid magma zone from where the other hybrid enclave globule types were also derived (e.g., Kumar and Rino, 2006), and hence they could be genetically linked (Zhang et al., 2014). These aspects need to be tested further using geochemical and isotopic investigations.